Skip to content
AI IntelligenceJul 19, 2026AI Intelligence
Article

A report shows that AI text detectors fail when language models can mimic an author's specific writing style.

For scientific texts, up to 48% of AI-generated passages went undetected. This calls the reliability of such verification tools into question.

Data Cube AI EditorialSource: The Decoder
01

Source Brief

A report shows that AI text detectors fail when language models can mimic an author's specific writing style. For scientific texts, up to 48% of AI-generated passages went undetected. This calls the reliability of such verification tools into question.

02